HRA-SFCP: Hybrid Resource-Aware SFC Placement in LEO Satellite Networks

With the rapid growth of satellite network technologies and emerging applications, Low Earth Orbit (LEO) satellite networks have become increasingly attractive for Service Function Chain (SFC) placement. However, the highly dynamic topology and constrained onboard resources in LEO satellite networks present substantial challenges for efficiently orchestrating SFC deployments. This paper introduces HRA-SFCP, a hybrid resource-aware SFC placement strategy. HRA-SFCP integrates a lightweight resource-efficient heuristic (RE-Heuristic) and a deep reinforcement learning framework (GT-PPO). An adaptive scheduler dynamically selects between the heuristic method and the DRL agent based on the available resource ratio: the heuristic is employed when resources are relatively abundant, while the DRL agent is invoked under resource constraints. Extensive experiments show that HRA-SFCP achieves a favorable trade-off between service acceptance and resource utilization, outperforming established baselines. These results validate HRASFCP as an effective and practical solution for SFC placement in dynamic LEO satellite environments.